Miscellaneous Concluding Observations. 191
The mean velocity deduced from these three observations
is, in round numbers, 9183 feet per second, the greatest
divergence from the mean being 10,353 feet for the highest
estimate and 7155 feet for the lowest estimate.
According to the results obtained by Milne," with more
precise time-data the velocities of propagation of an earthquake
along different lines vary from 4500 to 10,000 feet per second.
The mean result of 9183 feet is therefore a probable rough
approximation to the true mean velocity of the present
earthquake.
Although it is known that an earthquake travels from its
origin in different directions with different velocities, the
divergence in the results is far too great in the present case
to be attributable to differences in the conductivity of the
rocks along different lines, but is doubtless due to the errors
in the time-records. With this wide divergence it has ap-
peared to us useless to attempt to fix the origin by the
method of circles or hyperbolas, as adopted by Milne.
Observations on Direction.—Although during an earthquake
a particle at the surface of the earth may vibrate in many
azimuths (p. 33), the disturbance as a whole has a general
direction of propagation. If, therefore, the general direction
of the motion could be determined at a number of stations
round the epicentrum, the lines representing the directions
would intersect at the latter point.100 In the absence of
seismograph tracings, the only data available for such a
determination are the direction in which bodies are thrown
or moved, the direction of swing of suspended objects, and
the direction in which liquids are splashed. In making use
99 Trans. Seism. Soc. Japan, vol. vii., p. 11.
100 That is of course on the assumption that the general direction of
propagation coincides with the direction of oscillation of the normal
wave. It is by no means uncommon, however, for bodies at particular
stations to be thrown into oscillation in a direction at right angles to the
known direction of propagation, in which case it is probably the trans-
verse wave which produces the effect noticed. Milne has observed that
three instruments placed at the corners of a triangle, of which the sides
were 800 feet in length, give different diagrams of the same earthquake,
although when placed side by side their results agreed (Trans. Seism.
Soc. Japan, vol. vii., part 2, p. 5).